Machine-learning-based de-embedding method for radio-frequency circuit and device, and related device

By using a machine learning-based approach and a neural network prediction model, the problem of requiring a large number of calibration parts in existing high-frequency, high-precision de-embedding methods has been solved, achieving high precision and wide applicability while reducing testing costs.

WO2026237286A1PCT designated stage Publication Date: 2026-11-19SOUTH CHINA UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2025/133511
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-16
Filing Date
2025-11-07
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

In existing technologies, de-embedding methods require a large number of calibration components for high-frequency and high-precision testing, which increases processing costs and testing difficulty. Furthermore, they have poor versatility and cannot be applied to specific circuit structures and processes.

Method used

Using a machine learning-based approach, a dataset is generated by designing parameterized models of the test structure and the structure to be predicted, and a neural network prediction model is trained. High-frequency, high-precision de-embedding is achieved using a small number of calibration components, and it is applicable to a variety of circuit structures and processes.

Benefits of technology

It achieves high-frequency and high-precision de-embedding, reduces the number of calibration components, lowers processing and testing costs, and has wide applicability, not limited to specific circuit structures and processes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2025133511_19112026_PF_FP_ABST
    Figure CN2025133511_19112026_PF_FP_ABST
Patent Text Reader

Abstract

Disclosed in the present application are a machine-learning-based de-embedding method for a radio-frequency circuit and a device, and a related device. The method comprises: designing a test structure and a structure to be predicted; establishing parameterized models of the test structure and the structure to be predicted; performing a parameter sweep or Monte Carlo analysis on the established models of the test structure and the structure to be predicted, so as to generate a dataset; establishing a neural network prediction model, and using the generated dataset to train the model; inputting actually measured data of the test structure into the trained neural network prediction model, so as to output actually measured data of the structure to be predicted; and on the basis of the type of the structure to be predicted, processing the actually measured data of the test structure and the structure to be predicted, so as to obtain an actually measured result of a DUT. In the present application, a neural network model is trained to learn an intrinsic mapping relationship between data of a test structure and data of a structure to be predicted, and thus high-frequency and high-precision de-embedding can be realized by using a small number of calibration pieces; and the present application is widely applicable and is not limited to a specific circuit structure and process.
Need to check novelty before this filing date? Find Prior Art

Description

Machine learning-based de-embedding methods and related equipment for RF circuits and devices Technical Field

[0001] This application relates to the field of instrumentation and measurement, and more particularly to a machine learning-based de-embedding method and related equipment for radio frequency circuits and devices. Background Technology

[0002] Accurate de-embedding is crucial for achieving precise device and circuit modeling. This is because, in both board-level and on-chip circuit testing scenarios, the designed Device Under Test (DUT) is not directly connected to the test instrument's ports. Instead, an additional fixture circuit is added around the DUT to connect it to the test instrument. Therefore, during testing, the influence of the fixture needs to be removed to restore the true value of the DUT; this process is called de-embedding.

[0003] Conventional de-embedding methods mainly include equivalent circuit methods and cascade methods. The former primarily includes the open-circuit-short-circuit method and its derivatives, such as the open-circuit-short-circuit-load method and the open-open-short-circuit-short-circuit method. The latter mainly includes the through-circuit method and the TRL method. However, these two types of methods have several shortcomings. First, their de-embedding effect is closely related to the number of calibration components; the larger the de-embedding frequency range and the higher the accuracy, the more calibration components are required, which undoubtedly increases the processing cost of high-frequency, high-precision de-embedding. Second, more calibration components increase the testing difficulty, requiring testers to repeatedly prick the pins to ensure consistent test results, consuming a lot of time and increasing testing costs. Finally, the circuit equivalence assumptions made by conventional de-embedding methods regarding fixtures often only apply to specific circuit structures or processes, resulting in poor versatility. Summary of the Invention

[0004] To at least partially address one of the technical problems existing in the prior art, the purpose of this application is to provide a machine learning-based de-embedding method and related equipment for RF circuits and devices. This method combines high-frequency, high-precision de-embedding capabilities with strong versatility, while requiring only a small number of calibration components, significantly reducing processing and testing costs. Furthermore, this application has a wide range of applicability and is not limited to specific circuit structures and processes.

[0005] The first technical solution adopted in this application is:

[0006] A machine learning-based de-embedding method for radio frequency circuits and devices includes the following steps:

[0007] Design the test structure and the structure to be predicted;

[0008] Establish parameterized models for the test structure and the structure to be predicted;

[0009] Perform parameter scanning or Monte Carlo analysis on the established test structure and the structure to be predicted model to generate a dataset;

[0010] Build a neural network prediction model and train the model using the generated dataset;

[0011] The measured data of the test structure is input into the trained neural network prediction model to output the measured data of the structure to be predicted.

[0012] The measured data of the test structure and the structure to be predicted are processed according to the type of the structure to be predicted to obtain the measured results of the DUT.

[0013] Furthermore, the test structure includes not only a DUT test structure, but also one or more of a pass-through test structure, a line test structure, a load test structure, or a reflection test structure. For example, when the structure to be predicted is a DUT, typically only a DUT test structure is needed; when the structure to be predicted is a Fixture, typically only a DUT test structure and a pass-through test structure are needed.

[0014] Furthermore, the structure to be predicted is a Fixture or a DUT.

[0015] Furthermore, the parameterized models of the test structure and the structure to be predicted are one or more combinations of electromagnetic simulation models, equivalent circuit models, mathematical fitting models, or neural network models.

[0016] Furthermore, the generated dataset consists of test structure data samples and structure data samples to be predicted.

[0017] Further, the measured data of the test structure and the structure to be predicted are processed according to the type of the structure to be predicted, including:

[0018] When the structure to be predicted is a Fixture, the measured results of the DUT are calculated using the cascaded de-embedding formula: Among them, T DUT T LF T RF And T MEAS These represent the T-parameter matrices for the DUT, left Fixture, right Fixture, and DUT test structure, respectively.

[0019] When the structure to be predicted is a DUT, the output data of the neural network prediction model is the measured result of the DUT.

[0020] The second technical solution adopted in this application is:

[0021] A neural network prediction model is trained using the method described above. Its input is test structure data, and its output is the structure data to be predicted.

[0022] The third technical solution adopted in this application is:

[0023] A computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned machine learning-based de-embedding method for radio frequency circuits and devices.

[0024] The fourth technical solution adopted in this application is:

[0025] A computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement a machine learning-based de-embedding method for radio frequency circuits and devices as described above; or the storage medium storing the neural network prediction model described above.

[0026] The fifth technical solution adopted in this application is:

[0027] An electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement a machine learning-based de-embedding method for radio frequency circuits and devices as described above; or the storage medium stores the neural network prediction model described above.

[0028] The beneficial effects of this application are: This application provides a complete de-embedding scheme based on machine learning for radio frequency circuits and devices, covering dataset generation and neural network prediction model training. By training the neural network model to learn the inherent mapping relationship between test structure data and the structure data to be predicted, high-frequency and high-precision de-embedding can be achieved without relying on a large number of calibration components. Moreover, it has a wide range of applicable scenarios and is not limited to specific circuit structures and processes. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the following description of the relevant technical solutions in the embodiments of this application or the prior art is provided with accompanying drawings. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions in this application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0030] Figure 1 is a flowchart of the de-embedding method for radio frequency circuits and devices based on machine learning in Embodiment 1 of this application;

[0031] Figure 2 is a schematic diagram of the through-test structure and the Fixture to be predicted structure in Embodiment 2 of this application;

[0032] Figure 3 is a schematic diagram of the equivalent circuit topology of the through-test structure in Embodiment 2 of this application;

[0033] Figure 4 is a diagram showing the S-parameter fitting effect of the equivalent circuit model of the through-test structure in Embodiment 2 of this application.

[0034] Figure 5 is a graph showing the S-parameter fitting effect of the equivalent circuit model of the Fixture structure to be predicted in Embodiment 2 of this application.

[0035] Figure 6 is a schematic diagram of the dataset distribution generated by Monte Carlo analysis of the equivalent circuit model of the through-test structure in Embodiment 2 of this application;

[0036] Figure 7 is a prediction effect diagram of the neural network prediction model for the Fixture structure to be predicted in Embodiment 2 of this application, where the solid line is the model prediction result and the dashed line is the electromagnetic simulation result.

[0037] Figure 8 is a schematic diagram of the prediction accuracy of the neural network prediction model for the structure to be predicted in the Fixture in Embodiment 2 of this application;

[0038] Figure 9 is a schematic diagram of the pass-through test structure and the Fixture to be predicted structure in Embodiment 3 of this application;

[0039] Figure 10 is a schematic diagram of the network cascade topology of the pass-through test structure in Embodiment 3 of this application;

[0040] Figure 11 is a graph showing the S-parameter fitting effect of the polynomial model of the through-test structure in Embodiment 3 of this application;

[0041] Figure 12 is a graph showing the S-parameter fitting effect of the Fixture polynomial model of the structure to be predicted in Embodiment 3 of this application.

[0042] Figure 13 is a schematic diagram of the dataset distribution generated by Monte Carlo analysis of the equivalent circuit of the through-test structure in Embodiment 3 of this application;

[0043] Figure 14 is a prediction effect diagram of the neural network prediction model in Embodiment 3 of this application, where the solid line is the model prediction result and the dashed line is the electromagnetic simulation result.

[0044] Figure 15 is a schematic diagram of the prediction accuracy of the neural network prediction model in Embodiment 3 of this application. Detailed Implementation

[0045] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0046] The terminology used in the embodiments of this application is for the purpose of describing specific embodiments only and is not intended to limit the embodiments of this application. The singular forms "a," "described," and "the" used in the embodiments of this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. Furthermore, unless otherwise expressly limited, terms such as "set," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0047] In the description of this application, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0048] In the description of this application, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0049] In the description of this application, "and / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the related objects before and after it are in an "or" relationship.

[0050] Example 1:

[0051] As shown in Figure 1, this application provides a machine learning-based de-embedding method for radio frequency circuits and devices. By training a neural network model to learn the inherent mapping relationship between test data and the structure to be predicted, high-frequency, high-precision de-embedding can be achieved using a small number of calibration components. Furthermore, it has wide applicability and is not limited to specific circuit structures and processes. The method includes the following steps:

[0052] S1. Design the test structure and the structure to be predicted.

[0053] Specifically, the test structure may include, in addition to the DUT test structure, one or more of the following: pass-through test structure, line test structure, load test structure, and reflection test structure; for example, when the structure to be predicted is a DUT, only the DUT test structure is usually required; when the structure to be predicted is a Fixture, only the DUT test structure and the pass-through test structure are usually required.

[0054] The structure to be predicted is a Fixture or a DUT.

[0055] S2. Establish parameterized models for the test structure and the structure to be predicted.

[0056] Optionally, the parameterized models of the test structure and the structure to be predicted can be one or more combinations of electromagnetic simulation models, equivalent circuit models, mathematical fitting models, and neural network models.

[0057] S3. Perform parameter scanning or Monte Carlo analysis on the established test structure and the structure to be predicted model to generate a dataset.

[0058] Specifically, the generated dataset consists of test structure data samples and structure data samples to be predicted.

[0059] S4. Build a neural network prediction model and train the model using the generated dataset.

[0060] S5. Input the measured data of the test structure into the trained neural network prediction model to output the measured data of the structure to be predicted.

[0061] S6. Process the measured data of the test structure and the structure to be predicted according to the type of the structure to be predicted, and obtain the measured results of the DUT.

[0062] To obtain the actual measurement results of the test piece, the data processing is divided into two scenarios:

[0063] When the structure to be predicted is a Fixture, the measured results of the DUT are calculated using the cascaded de-embedding formula: Among them, T DUT T LF T RF And TMEAS The T-parameter matrices represent the DUT, left Fixture, right Fixture, and DUT test structure, respectively. When the structure to be predicted is the DUT, the output data of the neural network prediction model is the measured result of the DUT.

[0064] In one implementation method, this application selects a Fixture as the structure to be predicted, and the predicted data is the S-parameter frequency response, with a frequency range of 0.1–67 GHz. To verify the feasibility of the method, two prediction embodiments of the Fixture structure to be predicted are disclosed. Once the Fixture structure to be predicted is predicted, a cascaded de-embedding formula can be further employed. Calculate the measured results of the DUT, where T DUT T LF T RF And T MEAS The T-parameter matrices represent the DUT, left Fixture, right Fixture, and DUT test structure, respectively. Then, the T-parameters can be converted into S-parameters using the network parameter transformation formula.

[0065] Example 2:

[0066] Please refer to Figures 2-8. This embodiment discloses a prediction case for a transmission line type Fixture structure to be predicted. The test structure is a straight-through structure.

[0067] As shown in Figure 2, the transmission line type Fixture structure to be predicted is half of the through test structure. The transmission line Fixture has a linewidth W1 = 20 μm, a line length L1 = 240 μm, and a substrate of gallium arsenide with a dielectric constant of 12.9 and a substrate thickness of 75 μm.

[0068] In terms of parametric modeling, this embodiment establishes a lumped equivalent circuit model. Figure 3 shows the topology of the lumped equivalent circuit model of the through-test structure. i and z i The Π-shaped network is the equivalent circuit of the transmission line clamp, y em and z em The resulting Π-shaped network serves as the equivalent circuit for the additional electromagnetic field in the through-through test structure. Electromagnetic simulation is used to extract the network parameters of the additional electromagnetic field in the Fixture structure to be predicted and the through-through test structure. Vector fitting techniques are then used to establish y-values ​​for each. i z i y em and z em The lumped equivalent circuit model is shown in Figures 4 and 5. The established equivalent circuit model has a good fitting effect on both the through-test structure and the Fixture structure to be predicted.

[0069] In terms of dataset generation, this embodiment employs Monte Carlo analysis. After applying a + / -10% uniformly distributed Monte Carlo analysis to all parameters in the lumped equivalent circuit model, the distribution of the resulting 1001 through-test structure samples is shown in Figure 6. The deviation of these samples gradually increases with increasing frequency, consistent with actual physical phenomena.

[0070] In terms of the neural network prediction model, this embodiment uses a deep neural network. This deep neural network has 9 hidden layers, each of which is fully connected, and the activation function is the hyperbolic tangent function. The input of this deep neural network is the S-parameters of the direct-through test structure, and the output is the S-parameters of the fixture to be predicted structure. Before training, the dataset is first normalized to the interval [-1, 1]. During training, the Adam optimizer is used, with an initial learning rate of 0.001 and 10,000 epochs. Every 50 epochs, the learning rate is reduced to 0.98 times its original value. The validation set is generated by electromagnetic simulation parameter scanning. The W1 scan range is 18–22 μm with a step size of 0.5 μm, and the L1 scan range is 216–264 μm with a step size of 6 μm. Therefore, the validation set contains 81 samples.

[0071] Figures 7 and 8 respectively illustrate the S-parameter prediction effect and prediction accuracy of the Fixture structure to be predicted in this embodiment. The prediction accuracy is represented by the coefficient of determination; the closer the coefficient of determination is to 1, the higher the prediction accuracy. Considering that the transmission line Fixture is a left-right symmetrical structure, only S11 and S12 need to be verified. It can be seen that the prediction accuracy of both the real and imaginary parts of the two parts in this embodiment reaches above 0.99, achieving accurate prediction.

[0072] Example 3:

[0073] Please refer to Figures 9-15. This embodiment discloses a prediction case for a GSG-type Fixture structure to be predicted, and the test structure is a straight-through structure.

[0074] As shown in Figure 9, the GSG-type Fixture structure to be predicted is half of the through-test structure. The GSG-type Fixture has a linewidth W2 = 20 μm, a line length L2 = 120 μm, a signal pad width W3 = 60 μm, a ground pad width W4 = 80 μm, and uses gallium arsenide as the substrate with a dielectric constant of 12.9 and a substrate thickness of 75 μm.

[0075] In terms of parametric modeling, this embodiment establishes a polynomial mathematical fitting model. Figure 10 shows the network cascade topology of the pass-through test structure. Due to the discontinuity between the GSG pad and the metal interconnect, each Fixture consists of a cascaded GSG network (GSG), a discontinuity network (Dis.), and a metal interconnect network (Int.). The pass-through test structure is composed of a cascaded Left Fixture, an Extra EM in the pass-through test structure, and a Right Fixture. The network parameters of the GSG pad, discontinuity, metal interconnect, and extra EM are extracted using electromagnetic simulation, and polynomial coefficients are extracted using a polynomial fitting algorithm, such as the polynomial_fit() function in Python. As shown in Figures 11 and 12, the established polynomial mathematical fitting model has a good fitting effect on both the pass-through test structure and the Fixture to be predicted structure.

[0076] In terms of dataset generation, this embodiment employs Monte Carlo analysis. After applying a + / -10% uniformly distributed Monte Carlo analysis to all parameters in the lumped equivalent circuit model, the distribution of the resulting 1001 through-test structure samples is shown in Figure 13. The deviation of these samples gradually increases with increasing frequency, consistent with actual physical phenomena.

[0077] In terms of the neural network prediction model, this embodiment uses a one-dimensional convolutional neural network. This one-dimensional convolutional neural network has two layers: an input layer of one-dimensional convolutional layer and an output layer of fully connected layer. The activation function is the hyperbolic tangent function. The input of this deep neural network is the S-parameters of the direct test structure, and the output is the S-parameters of the fixed prediction structure. Before training, the dataset is first normalized to the interval [-1, 1]. During training, the Adam optimizer is used, with an initial learning rate of 0.001 and 10,000 epochs. Every 50 epochs, the learning rate is reduced to 0.98 times the original value. The validation set is generated by electromagnetic simulation parameter scanning. The scanning range of W2 is 18–22 μm with a step size of 0.5 μm, the scanning range of W3 is 55–65 μm with a step size of 5 μm, and the scanning range of W4 is 72–88 μm with a step size of 8 μm. L2 is fixed at 120 μm, so the validation set has a total of 81 samples.

[0078] Figures 14 and 15 respectively illustrate the S-parameter prediction effect and prediction accuracy of the Fixture structure to be predicted in this embodiment. The prediction accuracy is represented by the coefficient of determination; the closer the coefficient of determination is to 1, the higher the prediction accuracy. Considering that the GSG type Fixture is a non-symmetrical structure, it is necessary to verify S11, S12, and S22. It can be seen that the prediction accuracy of the real and imaginary parts of the three parameters in this embodiment reaches above 0.99, achieving accurate prediction.

[0079] In summary, the method presented in this application can achieve accurate fixture network parameter prediction with only a small number of calibration components (just one through-hole component). These fixture network parameters can then be used to further de-embedding the network parameters of the test structure. Furthermore, the method proposed in this application exhibits excellent robustness; the trained neural network prediction model can achieve accurate predictions even under large fluctuations in the physical parameters of both the test structure and the structure to be predicted. The machine learning-based de-embedding method for RF circuits and devices provided in this application combines high-frequency, high-precision de-embedding capability with strong versatility, while requiring only a small number of calibration components, significantly reducing manufacturing and testing costs.

[0080] Example 4:

[0081] A neural network prediction model is trained using the method described above. Its input is test structure data, and its output is the structure data to be predicted.

[0082] In this embodiment, the neural network prediction model can be implemented using deep neural networks, convolutional neural networks, or other neural networks. This neural network prediction model can achieve accurate predictions even under large fluctuations in the physical parameters of the test structure and the structure to be predicted, and has the same functions and technical effects as those in the above embodiments, which will not be elaborated upon here.

[0083] Example 5:

[0084] In some possible implementations, various aspects of the methods in the embodiments of this application can also be implemented as a program product, including program code. When the program product is run on a computer device, the program code causes the computer device to perform the steps of a machine learning-based de-embedding method for radio frequency circuits and devices according to various exemplary embodiments of this application described above. The executable computer program code or "code" used to perform the various embodiments can be written in high-level programming languages ​​such as C, C++, Python, Smalltalk, Java, JavaScript, Visual Basic, Structured Query Language (e.g., Transact-SQL), Perl, or in various other programming languages.

[0085] Example 6:

[0086] This application also provides a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set. The at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement a machine learning-based de-embedding method for radio frequency circuits and devices, as shown in FIG1.

[0087] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0088] Since the storage medium is the storage medium corresponding to a machine learning-based de-embedding method for radio frequency circuits and devices in this application embodiment, and the principle of the storage medium in solving the problem is similar to that of the method, the implementation of the storage medium can refer to the implementation process of the above method embodiment, and the repeated parts will not be described again.

[0089] Example 7:

[0090] This application also provides an electronic device, which includes a processor and a memory. The memory stores at least one instruction, at least one program, a code set, or an instruction set. The at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement a machine learning-based de-embedding method for radio frequency circuits and devices, as shown in FIG1.

[0091] It is understood that the memory may include random access memory (RAM) or read-only memory. Optionally, the memory may include non-transitory computer-readable storage medium. The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function, instructions for implementing the various method embodiments described above, etc.; the stored data area may store data created according to the use of the server, etc.

[0092] A processor may include one or more processing cores. The processor connects to various parts of the server via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in memory, and accessing data stored in memory to perform various server functions and process data. Optionally, the processor may be implemented using at least one of the following hardware forms: Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor may integrate one or more of the following: Central Processing Unit (CPU) and Modem. The CPU primarily handles the operating system and applications; the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip without being integrated into the processor.

[0093] Since this electronic device is the electronic device corresponding to the de-embedding method for radio frequency circuits and devices based on machine learning in the embodiments of this application, and the principle of solving the problem by this electronic device is similar to that of this method, the implementation of this electronic device can refer to the implementation process of the above method embodiments, and the repeated parts will not be described again.

[0094] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0095] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0096] The above embodiments are merely illustrative of the technical concept and features of this application, intended to enable those skilled in the art to understand the content of this application and implement it accordingly, and should not be construed as limiting the scope of protection of this application. All equivalent changes or modifications made based on the substance of the content of this application should be covered within the scope of protection of this application.

Claims

1. A machine learning-based de-embedding method for radio frequency circuits and devices, characterized in that, Includes the following steps: Design the test structure and the structure to be predicted; Establish parameterized models for the test structure and the structure to be predicted; Perform parameter scanning or Monte Carlo analysis on the established test structure and the structure to be predicted model to generate a dataset; Build a neural network prediction model and train the model using the generated dataset; The measured data of the test structure is input into the trained neural network prediction model to output the measured data of the structure to be predicted. The measured data of the test structure and the structure to be predicted are processed according to the type of the structure to be predicted to obtain the measured results of the DUT.

2. The de-embedding method for radio frequency circuits and devices based on machine learning according to claim 1, characterized in that, The test structure includes not only the DUT test structure, but also one or more of the following: pass-through test structure, line test structure, load test structure, or reflection test structure.

3. The de-embedding method for radio frequency circuits and devices based on machine learning according to claim 1, characterized in that, The structure to be predicted is a Fixture or a DUT.

4. The de-embedding method for radio frequency circuits and devices based on machine learning according to claim 1, characterized in that, The parameterized models of the test structure and the structure to be predicted are one or more combinations of electromagnetic simulation models, equivalent circuit models, mathematical fitting models, or neural network models.

5. The de-embedding method for radio frequency circuits and devices based on machine learning according to claim 1, characterized in that, The generated dataset consists of test structure data samples and structure data samples to be predicted.

6. The de-embedding method for radio frequency circuits and devices based on machine learning according to claim 1, characterized in that, The measured data of the test structure and the structure to be predicted are processed according to the type of the structure to be predicted, including: When the structure to be predicted is a Fixture, the measured results of the DUT are calculated using the cascaded de-embedding formula: Among them, T DUT T LF T RF And T MEAS These represent the T-parameter matrices for the DUT, left Fixture, right Fixture, and DUT test structure, respectively. When the structure to be predicted is a DUT, the output data of the neural network prediction model is the measured result of the DUT.

7. A neural network prediction model, characterized in that, The data is trained by the method described in any one of claims 1-5, with the input being test structure data and the output being the structure data to be predicted.

8. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed by a processor, are used to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The storage medium stores at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the method as described in any one of claims 1 to 6; or the storage medium stores the neural network prediction model as described in claim 7.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the method as described in any one of claims 1 to 6; or, the memory stores the neural network prediction model as described in claim 7.